Estimation risk occurs when parameters relevant for decision making are uncertain. Bayes' criterion is consistent with expected-utility maximization in the presence of estimation risk. This article examines optimal (Bayes') land allocations and land allocations obtained using the traditional plug-in approach and two alternative decision rules. Bayes' allocations are much better economically than the other allocations when there are few sample observations relative to activities. Calculation of certainty equivalent returns (CERs) with estimation risk is also discussed and illustrated. CERs are typically (and incorrectly) calculated with the plug-in approach. Plug-in CERs may be extremely misleading.
Available at: http://works.bepress.com/dermot_hayes/14/